Combined Shear and Tension Stress 3

[RFreund]

This is probably a lay-up for someone. However I'm having a brain-shart. If you want to check a plate for combined shear and tension how should the stresses be combined?
For example a pair of plates connecting a beam to a column for a fixed for torsion and moment end condition. Is the combined stress check for the plate fv/Fv + ft/Ft ? or should they be squared? Also is the required force on the weld at the column and plate interface equal to F=sqrt(T^2+V^2) or should it just be the sum of the magnitudes?
Is this referenced in AISC or Salmon and Johnson...I feel like I have read more on this than I am recalling.

Thanks again

EIT

 

[rb1957]

combine them however you want to ! depending on how conservative you want to be ...

simple addition (v/V+t/T) is most conservative
sum of squares is probably more common at is similar to principal stress.

note, your question is very similar to another thread, combining shear and bending stresses; that might give you some insight.

 

[ToadJones]

If you are designing a structural steel connection, then may I direct you to bottom of page 9-3 of AISC 360-05; "Connection Elements Subject to Combined Loading"

 

[RFreund]

rb - I did read the other thread and that is basically what got this started and the fact that I have a connection design similar to the situation previously described. Why do you say that it is similar to the principle stresses?

Toad - Thanks I figured that equation would come into play, but I'm still not sure how it applies or how to make the assumptions.

Would it be correct to size the weld based on the sum of the squares?

EIT

 

[ToadJones]

what equation?
Read closer.

as for the weld, calculate the maximum resultant stress on the point of the weld where max stress occurs (vectorally). Design the weld for this stress.
If you have forces, say 90º to one another, then, yes, use sum of the squares. But say you have tensile bending stress at the same location and a direct axial tension stress, then no, these do not combine as the sum of the squares...they would be additive, and then possible combined with other stresses vectorally.
Sketch your forces, you will see.

 

[rb1957]

if you had a tension stress in one direction (say x-) and zero in y- and a shear stress, then the principAL stress is the sum of the squares.

 

[PSlem]

If significant torsional load it will be at 45*. Proportions of stresses will determine critical angle.

 

[RFreund]

Ahh, yes. Thanks for the comments.

rb - sorry I responded with out really reading and thinking through your post.

Toad - Actually I was referring to the Von-Mises criterion equation. However I guess this is actually just showing me that combining the stresses is not adding any accuracy to the design.
And after rereading the page it does make sense now. If I know my normal stress caused by resolving the moment in form of uniform tension on the plate and shear stress from the torsional moment, I still do not know what the normal stress is in the perpendicular direction.


EIT

 

[hokie66]

What's a brain-shart? Same as a brain fart?

 

[BAretired]

A sketch might help.

BA

 

[SAIL3]

Discovering that Von-Mises stress analysis that has been used and accepted for decades is no longer valid/accurate is a spectacular revelation.
This breakthrough in knowledge, alone, justifies the publication of the AISC 13th ed, in my opinion.

 

[ToadJones]

Sail-
I'm not sure if you are throwing wholesale comments at the post or if you are being genuine.
Von Mises is widely used and accepted, yes.
This doesn't mean it applies to every situation without exception. If I were designing a pressure vessel like an industrial flue I would use Von Mises.

It would be great if ConnectEgr could chime in here. This is in his wheelhouse.

I'm inclined to think, that if AISC says connections (small/compact) need not be designed for combined stresses, then one does need to.

I do check for combined stress in these situations at times using a generic interaction equation.
If you check a connection design and see that conservatively you require a 1/2" gusset plate, then why knit pick the hell out of it?

Connect?

 

[rb1957]

"I still do not know what the normal stress is in the perpendicular direction" ... why wouldn't it be due to poission effect ? same as in the transverse direction.

so really you know how to combine stresses together (using von mises), you just don't think you have the complete definition of the state of stress ??

 

[RFreund]

hokie - I'm glad you asked (snicker). A brain-shart is very similar to a brain fart difference being in the shart part. See a shart is not a pleasant experience and is usually not done purposely. Let just say its a fart that leaves some substance in your shorts.

Back to the matter at hand - slightly modified as I think I may need to do some stress/strain analysis studying.

rb - I will need to review Rorak and stress transformation equations on Monday however how would you use poission to determine the stress? I can imagine using poission for a member in uniform axial stress. But like I said I probably need to do some studying.

See the attached paper provided by WillsV earlier which explains my original question. However I still would like to know if it is possible to find the stress perpendicular to the section cut under consideration (sigma-y). Or if any one has an example in designing an element using Von-mises that would be great.

Thanks again.

EIT

 

[RFreund]

forgot to attach

EIT

 

[dhengr]

Rfreund:

A sketch of your actual connection would be real helpful, otherwise your OP could probably describe at least several dozen slightly different details; all of which might be treated slightly differently, in their stress analysis. There are some very good trains of thought in the other thread, which you participated in briefly. I am at a distinct disadvantage in some of these discussions because people keep referring. to a page and formula number in the latest ed. of codes which I may not have a copy of. I do wish people would give the formula or quote the para., but that may be much more than I have a right to expect; I could buy the newest code too if it would make any money for me, instead of just cost me to play here.

I believe Lion’s main idea, in the other thread, was to design one part of the joint for the shear loads, and one part for the tension or bending loads and then let them coexist. They will share all the loads on the joint, and exactly how, we might never know. WillisV suggests that there may be a few cases on connection details where we might want to consider combined stresses, and I agree with him, then the trick is, when and where. But, then Thornton’s paper suggests this may be beating a dead horse, in most cases.

I keep harping on simple designs and clean details, and at the same time you should have a good feel for where some of these max. combined stress may occur, and knowing that at one corner they may be additive and at the opp. corner they may lead to a reduced max., but maybe never knowing their exact magnitude. Even with all the fancy computer software we have today our analysis is not an exact science, except on the very simplest of structures and connections, or you could spend the rest of your life designing a couple connections, maybe more exacter..., or not. Otherwise, the most important thing might be knowing where these higher stresses can occur, and detailing and welding in those areas to prevent any poor quality, stress raisers or hard spots which could lead to failure. Strain rate could come into play in a few instances like EQ events, and maybe fatigue in some cases, but most std. structural connections and stresses don’t fit these conditions. How you reinf. and terminate the end of a weld line where these higher stresses might occur may be more important than the exact stress magnitude, if you could pin it down. Certainly, you don’t want craters or undercuts at the ends of welds. You might provide a partial bevel at the last few inches, so that you end up with the typical fillet reinforcing the bevel and lowering the weld stress in that area. Be very careful about welding around sharp corners which almost always leaves an undercut notch in the corner material. Attention to these kinds of details may be more important than knowing the max. combined stress to some uncertain degree. Which is the biggest stress on the weld, worry most about that, how does it mistreat your weld?

Everyone is all worried about VonMise, but few people seem to know who he is or what he does, or why or if he is important. If, through your detailing, you are careful about where and how a little yielding or plasticity happens, and it is well protected by material, all around it, which does not reach yield, this is seldom a failure mechanism, and probably happens far more than we actually imagine. High stresses and the wrong orientation at the root of a weld can be disastrous. For the most part, I’ve looked at combined stresses and max. stresses when faced with fatigue problems, not static problems.

 

[rb1957]

"if it is possible to find the stress perpendicular to the section cut under consideration" ... only by analyzing your particular structure. it could be, as your ref notes, 0, +ve, or -ve ... but anyone knows without looking at your structure.

 

[RFreund]

I will post a sketch in a bit. Currently the power is out( on my phone now). Thanks though for the replies.

EIT

 

[jrw501]

Von Mises would suggest (ft/Ft)^2+(4/3)(fv/Fv)^2=1 (assuming a safety factor of 1.67), but this is conservative for low tension stresses. For reference you may want to take a look at a paper by Subhash Goel "Combined Shear and Tension Stresses".

 

[rb1957]

not sure i get that ...
from von mises, Ft^2 = (ft^2+3fv^2)*SF^2,
SF = Ft/sqrt(ft^2+3fv^2) = Ft/fvm

no?